Diagnosing Errors in DbC Programs Using Constraint Programming

Abstract

Model-Based Diagnosis allows to determine why a correctly designed system does not work as it was expected. In this paper, we propose a methodology for software diagnosis which is based on the combination of Design by Contract, Model-Based Diagnosis and Constraint Programming. The contracts are specified by assertions embedded in the source code. These assertions and an abstraction of the source code are transformed into constraints, in order to obtain the model of the system. Afterwards, a goal function is created for detecting which assertions or source code statements are incorrect. The application of this methodology is automatic and is based on Constraint Programming techniques. The originality of this work stems from the transformation of contracts and source code into constraints, in order to determine which assertions and source code statements are not consistent with the specification.

On the Use of Constraints in Dynamic Slicing for Program Debugging

“…Note that this step distinguishes our approach from the previously published approaches, which also make use of constraints. In [26], [27], [31], [32], [33], [34] the whole program and not the execution trace is converted. This leads to a larger constraint system and imposes additional assumptions like the considered maximum number of iterations.…”

“…We call such a program part an explanation. There are many approaches that are capable of returning explanations including [21], [22], [23], [24], [25] and [26], [27] among others.…”

“…There is also work describing the application of constraints for the same purpose, e.g., [26], [27], [32] among others. In contrast to these papers we are not relying on a model capturing all possible paths through a program.…”

Generating Distinguishing Tests Using the Minion Constraint Solver

“…e = e -1; } 6. return res; As loops cannot be directly converted to our constraint representation, past works have proposed to unroll loops, i.e., to create a loop-free program by replacing the loop of the original program by a set of nested if-statements (e.g., see [2] and [13]). If the maximum number of iterations is known in advance, then the loop-free program is equivalent to the original program.…”

“…Constraints have been used for various purposes like verification [4], debugging [2], [13], program understanding [14] as well as testing [6], [9], [10]. Some of the proposed techniques use constraints to state specification knowledge like pre-and post-conditions.…”